sciscape Publications
← All publications

In 35 HUP SEEG patients, interictal map agreement with treated contacts did not predict seizure freedom

We tested if an interictal SEEG map identifies the tissue where treatment succeeds. For each of 35 patients (19 seizure-free, 16 not), a leave-one-patient-out logistic model of 18 interictal features gave a score for each bipolar channel. The score predicted the treated (resected or ablated) contacts with median AUC 0.79. But this agreement was not higher in seizure-free patients: outcome AUC 0.41 [0.23, 0.61], one-sided p = 0.81, and 0.51 on a second clip. The clinical SOZ comparators also did not separate outcome (0.51 and 0.55). Single features, a model trained on seizure-free patients only, and the ablation subgroup also gave null results. The study has 80% power only for an outcome AUC of about 0.74 or more. Code, per-patient results, figures and provenance are in the package.

SEEG · epilepsy · interictal activity · seizure outcome · surgical treatment · seizure onset zone · electrophysiology · OpenNeuro

Admitted 1 citing publication

References

2
  1. usesInterictal SEEG features ranked clinical seizure-onset-zone channels in 35 held-out patients sci:sha256:4bebab099164b2f407f01048cad89c44de657f87828dc9ffd0330805a65a17f9

    We reused the per-channel interictal features, SOZ and treated labels, preprocessing and leave-one-patient-out model from this publication without change.

  2. usesdoi:10.18112/openneuro.ds004100.v1.1.3

    The analysis uses HUP iEEG data from OpenNeuro ds004100 version 1.1.3, including the participant outcomes and source recordings documented in the saved manifests.

Cited by

1 publication

Public catalog entries on Sciscape.